Not random, but not predictable. That's the essence of complexity or chaos theory, and it's popularly captured by the idea of a butterfly flapping its wings on one continent setting off a cyclone on another. Compare this to the old Newtonian conception of nature where every action is followed by an equal and opposite reaction.
I've been following the 'debate' over climate change and have been surprised at some of the arguments put forward by self-styled skeptics. I've seen a lot of people rejecting climate change because while CO2 gas continues to be emitted into the atmosphere, temperatures over the past 15 years or so have not appreciably increased. The very same people often also reject climate change because of the failure of the computer models to predict this 15 year hiatus. I don't take this objection very seriously because using computers to accurately predict the weather or climate is probably one of the most mind numbingly difficult intellectual challenges we have ever faced, and the failure of computers to accurately predict the weather or the climate is nothing new; it's been the norm ever since the problem was first taken on. To reject climate change out of thwarted expectations that computer models would suddenly give accurate predictions seems either naive or disingenuous.
The other objection, that we've been continually emitting billions of tonnes of CO2 into the atmosphere without any appreciable increase in temperature over the past 15 years, is more interesting. I think it comes from a linear, Newtonian mind set that underestimates the complexity of the climate and the physics of the atmosphere. It's intuitively satisfying that every billion tonnes of gas emitted should raise the temperature by some fraction of a degree, in some kind of lock step, action/reaction fashion, but our double pendulum showed us that intuition will only take you so far, and perhaps in the wrong direction.